On Aug. 14, Bank of America research said the AI-driven rise in U.S. stocks is starting to show hallmarks of the late-stage 1999 internet bubble. The clearest signs are unusually high S&P 500 concentration, record-low active stock-picking (fund managers choosing individual shares), and continued investor demand for a small group of AI winners. The warning has intensified after Nvidia and Wall Street giants began planning $500 billion in AI infrastructure financing. Supporters see the structure as a longer-term funding channel for AI customers that could sustain demand for chips, data centers and computing capacity, while skeptics say it resembles vendor financing (supplier-backed customer funding) and could turn AI capital expenditure into a more complex financial arrangement. Bank of America argued that when chipmakers, cloud providers and large technology companies drive most of an index's gains, headline strength can mask underlying fragility. If passive inflows (index-tracking fund money) keep concentrating in the largest AI stocks while active managers step back, the rally becomes more dependent on a few companies meeting earnings expectations. If AI revenue grows more slowly than capital spending, index volatility could increase.